For the first 5 games of the season, every team has a pre-season ranking. (Without this ranking, there’s no difference in + or – points when beating NJIT or UMD, no difference in + or – points when losing to Denver or to VMI. And, as a consequence, the points are erratic and, more importantly, take more than a season to adjust itself. So we have a pre-season ranking.) At 6 Weeks, whatever the current points are for the releant team, that becomes their new “pre-season ranking”.
But Why Six Weeks?
Because we want to make a transition, as soon as possible, from having any connection to the previous season to having a ranking that is based exclusively on the current season. This is what we learned by playing with our spreadsheet:
If NJIT (with 310 points) was actually as good as Albany (with 980 points) and, as a consequence, obviously mis-ranked by placing them at the bottom of the list with 310 points, based on teams receiving more than 10% of the losing teams when the losing team is higher rated than the winning team and the greater the disparity the greater the number of points, it will take less than five weeks for NJIT to grow to 1,500 points.
If a team, like Furman, has a pre-season rank of 600 points, a number closer to 980 but not so close that it would be simple to accumulate 1,500 points, is actually mis-ranked and, in reality, like Albany, should have 980 points, it takes five weeks for this team to attain 1,500 points.
As a consequence, in order to cover any possible scenarios, unlikely as they may be, we wait five weeks before adjusting every teams “pre-season ranking”.
